摘要
在平台式惯导系统中可通过阻尼网络来阻尼系统的振荡误差,而在捷联惯性航姿系统中也可以借鉴这个思想,通过加速度计的信息对陀螺漂移进行修正.文中提出了分别利用陀螺仪和加速度计来解算载体姿态,然后通过卡尔曼滤波器进行数据融合的方法.由于加速度计信息进行阻尼的方法只有在系统处于非加速度运动状态才可用,因此,本文将故障检测理论中的残差χ2检验法应用在卡尔曼滤波器中,进行运动状态的实时判断,并根据判断结果进行相应的处理.此外,算法中还考虑到传感器测量野值带来的影响,提高了算法的容错性能和灵敏度.最后,采用实际系统的动静态试验验证了本算法的有效性.
In a platform inertial navigation system, damping network is applied to damp the gyro＇s drift. For strap-down inertial attitude and heading reference system （SIAHRS）, gyro＇ s drift error can also be suppressed by the information from accelerometers. This paper proposes the idea of calculating vehicle attitude by both gyros and accelerometers, and designs a damping Kalman filter to integrate this information. However, only when the vehicle was moving without acceleration can one use the output of accelerometers to calculate its attitude. For this reason, the residual chi-square test is employed in the Kalman filter to detect vehicle movement in real-time. Besides, effects of outliers from measuring sensors are also considered in order to enhance system＇ s fault tolerant ability and sensitivity. Results of static and dynamic experiments are carried out.
出处
《应用科学学报》
CAS
CSCD
北大核心
2007年第2期183-188,共6页
Journal of Applied Sciences
